owner stringclasses 14
values | repo stringclasses 14
values | id int64 116k 4.61B | issue_number int32 1 181k | author stringlengths 1 39 | body stringlengths 1 262k | created_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:16:18 | updated_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:22:56 | reactions unknown | author_association stringclasses 5
values |
|---|---|---|---|---|---|---|---|---|---|
ClickHouse | ClickHouse | 657,469,756 | 12,457 | 4ertus2 | > hashed с key UInt32 вообще бы не должен создаваться | 2020-07-13T10:16:12 | 2020-07-13T10:16:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,951,073 | 12,455 | amosbird | After futher thoughts, two special Field types are added (-Inf and +Inf) which explicitly helps special comparisons such as NULL_LAST. Field Null should not be used as comparison argument. | 2020-07-14T03:53:52 | 2020-07-14T03:53:52 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 658,001,347 | 12,443 | alexey-milovidov | Ok, probably it is because I did not merge with master and performance of master was improved.
@akuzm Why it is not merged with master before running tests? | 2020-07-14T06:44:39 | 2020-07-14T06:44:39 | {} | MEMBER |
ClickHouse | ClickHouse | 658,002,741 | 12,424 | alexey-milovidov | @akuzm
> Will do, but it's probably going to be too noisy.
When we had our old performance tests, average difference larger than 2% have always indicated significant changes. | 2020-07-14T06:48:30 | 2020-07-14T06:48:30 | {} | MEMBER |
ClickHouse | ClickHouse | 658,053,473 | 12,342 | alexey-milovidov | It's first appeared in 20.5, it means that we can safely rollback the change. | 2020-07-14T08:41:40 | 2020-07-14T08:41:40 | {} | MEMBER |
ClickHouse | ClickHouse | 657,324,971 | 12,435 | l1t1 | I extract the server rpm file with 7z tools, and use following command,which includes the config.xml extracted from the rpm , then the server runs
```
/usr/bin/clickhouse server --config=./etc/clickhouse-server/config.xml
`` | 2020-07-13T02:25:37 | 2020-07-13T02:25:37 | {} | NONE |
ClickHouse | ClickHouse | 657,369,333 | 12,330 | alexey-milovidov | It's just non-conforming C++ code in `libhdfs3`:
Calling
```
&buffer[0]
```
on empty vector is undefined behaviour.
It does not lead to any bugs.
The library is no longer supported and we have to fix it by ourself.
The fix is trivial. | 2020-07-13T05:38:06 | 2020-07-13T05:38:06 | {} | MEMBER |
ClickHouse | ClickHouse | 657,641,415 | 12,438 | akuzm | > @akuzm The result of performance test is not published.
It's probably the S3 quota troubles, and also the fact that the CI is overloaded and the perf test tasks wait in queue for more than a day sometimes. There is a workaround to get it sooner -- click 'Details' and increase the task priority manually (i've done ... | 2020-07-13T15:52:55 | 2020-07-13T15:52:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,999,930 | 12,478 | alexey-milovidov | Allow to specify full query to pass to external data source can be Ok but a few complications exist:
- sometimes the query will not parse by ClickHouse and we will have to put in in quotes (and escaping will be cumbersome);
- we have to derive data types of resulting columns by sending another request. | 2020-07-14T06:40:45 | 2020-07-14T06:40:45 | {} | MEMBER |
ClickHouse | ClickHouse | 658,044,404 | 10,561 | alesapin | ubsan
```../contrib/lz4/lib/lz4.c:818:44: runtime error: applying zero offset to null pointer```
Missing a patch in lz4? | 2020-07-14T08:24:17 | 2020-07-14T08:24:17 | {} | MEMBER |
ClickHouse | ClickHouse | 658,046,469 | 12,386 | alexey-milovidov | Another question - do you feel that it's necessary to have a separate name for this storage / table function?
Maybe if `s3` will just magically support COS is also Ok? | 2020-07-14T08:28:17 | 2020-07-14T08:28:17 | {} | MEMBER |
ClickHouse | ClickHouse | 657,236,552 | 12,434 | excitoon | <s>
```
SELECT
DayOfWeek,
count(*) AS c
FROM datasets.ontime_wide
WHERE (Year >= 2000) AND (Year <= 2008)
GROUP BY DayOfWeek
ORDER BY c DESC
↘ Progress: 58.95 million rows, 176.85 MB (48.78 million rows/s., 146.34 MB/s.) ███████████████████████████████████████████████████████████████████▌ 98%
Rec... | 2020-07-12T15:22:49 | 2020-07-12T21:57:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,282,624 | 12,424 | azat | >The problem is in ParserQueryWithOutput, when it is used (i.e. when FORMAT is passed) then SETTINGS after FORMAT are not applied:
AFAIU this is the intention, but looks tricky, maybe worth adding them to the ASTSelectQuery too after parsing from ParserQueryWithOutput? | 2020-07-12T22:09:37 | 2020-07-12T22:09:37 | {} | MEMBER |
ClickHouse | ClickHouse | 657,349,957 | 12,435 | l1t1 | after installed the rpm from http://www.rpmfind.net/linux/centos/8-stream/BaseOS/x86_64/os/Packages/initscripts-10.00.7-1.el8.x86_64.rpm, the clickhouse server rpm can be installed now,
```
[root@b32bc29e6c28 tmp]# rpm -ivh clickhouse-server-20.6.1.4050-2.noarch.rpm
warning: clickhouse-server-20.6.1.4050-2.noarch.rp... | 2020-07-13T04:20:22 | 2020-07-13T04:20:22 | {} | NONE |
ClickHouse | ClickHouse | 657,438,507 | 12,454 | MrSsunlight | If I move `/var/lib/clickhouse/metadata/dnsmon/rtypes.sql /tmp/` before starting clickhouse, clickhouse can start successfully；And, after the startup is successful, move the file back to the source path, attach the table in clickhouse-client, it can be successfully loaded, no exception information
```
[root@826f2ed... | 2020-07-13T09:24:59 | 2020-07-13T09:28:35 | {} | NONE |
ClickHouse | ClickHouse | 607,134,350 | 4,006 | zhang2014 | Update:
Maybe we can :
``` sql
CREATE DATABASE clickhouse_database_name ENGINE = MySQL(...) SETTINGS locality = true
[, default_table_engine = CollapsingMergeTree]
```
`MySQL` tables will be created in `ClickHouse`(table engine = `CollapsingMergeTree`), And the data will be obtained through the [MySQL Replicati... | 2020-04-01T09:15:11 | 2020-07-13T10:45:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,639,510 | 12,278 | akuzm | > > Performance — 1 faster, 7 slower
>
> Now totally different set of tests become slower.
I'm looking at these reports:
https://clickhouse-test-reports.s3.yandex.net/12278/8092f9f8618955e2a4311149acaaa2edc1f82b3a/performance_comparison/report.html#fail1
https://clickhouse-test-reports.s3.yandex.net/12278/edb... | 2020-07-13T15:49:23 | 2020-07-13T15:49:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,288,140 | 8,941 | alexey-milovidov | > There is a alternative solution for my particular case - to implement DateTimeOffset data type, which additionally stores the offset component for the concrete timezone, but this is a separate much more complicated task.
Yes, you can store time zone offset. But care should be taken because single time zone may hav... | 2020-07-12T23:01:33 | 2020-07-12T23:01:33 | {} | MEMBER |
ClickHouse | ClickHouse | 657,294,449 | 12,434 | excitoon | FYI @alexey-milovidov The same optimization can also improve performance of spinning disks. | 2020-07-12T23:51:05 | 2020-07-12T23:51:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,420,108 | 11,536 | vitlibar | I've made another PR https://github.com/ClickHouse/ClickHouse/pull/12423 to continue the work. | 2020-07-13T08:54:36 | 2020-07-13T08:54:56 | {} | MEMBER |
ClickHouse | ClickHouse | 657,553,721 | 9,624 | alesapin | All important tests should be fixed, but flaky failures still may happen. | 2020-07-13T13:15:17 | 2020-07-13T13:15:17 | {} | MEMBER |
ClickHouse | ClickHouse | 658,010,033 | 12,126 | alexey-milovidov | See also https://github.com/GoogleContainerTools/distroless | 2020-07-14T07:06:57 | 2020-07-14T07:06:57 | {} | MEMBER |
ClickHouse | ClickHouse | 658,044,640 | 10,561 | alesapin | DatabaseAtomic failures not important, experimental functionality. | 2020-07-14T08:24:44 | 2020-07-14T08:24:44 | {} | MEMBER |
ClickHouse | ClickHouse | 658,046,649 | 10,561 | alesapin | No integration tests. Run them manually. | 2020-07-14T08:28:38 | 2020-07-14T08:28:38 | {} | MEMBER |
ClickHouse | ClickHouse | 657,277,160 | 11,815 | alexey-milovidov | Hyperscan and async unwinding: https://clickhouse-test-reports.s3.yandex.net/12440/e8a3f0abb3e3169f0ade88bc2edd076ea4a0cff1/functional_stateful_tests_(release).html#fail1 | 2020-07-12T21:22:51 | 2020-07-12T21:22:51 | {} | MEMBER |
ClickHouse | ClickHouse | 657,405,029 | 10,994 | MrSsunlight | > @vitlibar @alexey-milovidov it's impossible to start CH 20.4+ if one have a table with Engine=Dictionary (old xml dictionary).
>
> to start CH, it needs to remove tables metadata ( `rm /var/lib/clickhouse/metadata/default/node.sql` ) start CH, create tables.
Is this feature newly added in version 20.x? When I r... | 2020-07-13T08:27:21 | 2020-07-13T08:27:21 | {} | NONE |
ClickHouse | ClickHouse | 657,596,192 | 12,424 | akuzm | > @akuzm Could you please provide a number - what is "overall" (aggregated) change in performance across all tests?
> (usually it is geomean across all queries, but geomean of geomeans of queries of tests is probably better).
Will do, but it's probably going to be too noisy. I liked the cumulative distribution func... | 2020-07-13T14:32:20 | 2020-07-14T06:46:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,287,702 | 8,941 | alexey-milovidov | DateTime is stored as a number containing unix timestamp (the number of seconds from 1970-01-01 00:00:00 UTC).
To convert text (calendar) representation to unix timestamp and vice-versa, you have to use some timezone (probably UTC). But you cannot avoid using timezone.
> 1. toDateTime('2020-01-01 12:00:00', 'Unsp... | 2020-07-12T22:57:13 | 2020-07-12T22:57:13 | {
"-1": 1
} | MEMBER |
ClickHouse | ClickHouse | 657,277,072 | 12,440 | alexey-milovidov | > 00095_hyperscan_profiler
Looks like Hyperscan still has some issues with async unwinding.
(reported) | 2020-07-12T21:22:13 | 2020-07-12T23:50:11 | {} | MEMBER |
ClickHouse | ClickHouse | 657,313,743 | 12,448 | alexey-milovidov | ```
grep -F '"command":' compile_commands.json | grep -v -P ' -c .+?/contrib/' | grep -o -P ' (-Wno-\S+|-w )' | sort | uniq -c | sort -rn | awk '{ if ($1 >= 1000) print $2 }' | sort
``` | 2020-07-13T01:33:32 | 2020-07-13T01:33:32 | {} | MEMBER |
ClickHouse | ClickHouse | 657,328,000 | 12,424 | alexey-milovidov | @akuzm Could you please provide a number - what is "overall" (aggregated) change in performance across all tests?
(usually it is geomean across all queries, but geomean of geomeans of queries of tests is probably better). | 2020-07-13T02:39:29 | 2020-07-13T02:39:29 | {} | MEMBER |
ClickHouse | ClickHouse | 657,357,864 | 12,081 | alexey-milovidov | @flipper86 We have `JSONAsString` format in version 20.5.
This format reads JSONs to a table with single String field containing whole JSON object.
You can use this format to parse JSON as String and then use JSONExtract... functions to extract fields from nested data. | 2020-07-13T04:56:17 | 2020-07-13T05:00:28 | {} | MEMBER |
ClickHouse | ClickHouse | 657,471,462 | 12,450 | YiuRULE | @alesapin Yep, here the complete stacktrace
```
Jul 13 12:04:35 nightly clickhouse-server[3757]: 2020.07.13 10:04:35.541800 [ 3778 ] {} <Error> void DB::BackgroundProcessingPool::threadFunction(): Code: 424, e.displayText() = DB::ErrnoException: Destination file /srv/clickhouse/data/default/test/tmp_mut_20200318_1936... | 2020-07-13T10:19:29 | 2020-07-13T10:19:29 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,611,288 | 12,078 | Akazz | It is worth noting that this behaviour does not reproduce with CH v20.5. What's also interesting is that file `src/Processors/Formats/Impl/CSVRowInputFormat.cpp` did not change since v20.4. I will continue looking into it | 2020-07-13T14:59:08 | 2020-07-13T14:59:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,623,253 | 12,427 | akuzm | Test failures are due to internal CI problem (quota limit in S3 reached). | 2020-07-13T15:20:11 | 2020-07-13T15:20:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,675,920 | 12,428 | alexey-milovidov | This is better: #12443 | 2020-07-13T16:59:02 | 2020-07-13T16:59:02 | {} | MEMBER |
ClickHouse | ClickHouse | 657,988,583 | 6,943 | hzruandd | > @alexey-milovidov @filimonov We limit query memory limit to 4GB and encounter memory limit exception when inserting into tables with more than 2000 columns. Our current workaround is to create multiple tables to reduce columns per table.
>
> It would be nice to reduce pre-allocated column buffer size when number o... | 2020-07-14T06:09:28 | 2020-07-14T06:09:28 | {} | NONE |
ClickHouse | ClickHouse | 657,372,925 | 12,330 | alexey-milovidov | Note that most of C and C++ libraries are not UBSan clean.
Actually we are one of the few in "elite club" of developers who continuously test their code with ASan+TSan+MSan+UBSan+static analysis+whatever.
But we have some 3rd-party libraries with disabled UBSan instrumentation and it is normal. | 2020-07-13T05:50:17 | 2020-07-13T05:51:30 | {} | MEMBER |
ClickHouse | ClickHouse | 657,580,665 | 11,709 | 4ertus2 | Continued here #12467 | 2020-07-13T14:03:56 | 2020-07-13T14:03:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,700,078 | 12,448 | alexey-milovidov | The most convenient it to create a new branch, then pull all commits from this PR, then add more commits. | 2020-07-13T17:48:05 | 2020-07-13T17:48:05 | {} | MEMBER |
ClickHouse | ClickHouse | 657,731,189 | 12,474 | den-crane | 1 select where in (select) / select where not in
2 https://gist.github.com/alexey-milovidov/2ebeb97a6dfe2e885e6cbae3bb2d7545#file-rank-txt-L63
```
select empid, deptid, salary, salaries[1] salary_max_over_dep, rank
from (
SELECT deptid, groupArray(empid) AS empids, groupArray(salary) AS salaries
FROM (... | 2020-07-13T18:52:26 | 2020-07-13T18:52:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,045,890 | 12,386 | alexey-milovidov | Ok. Now we have to fix Style check and Build checks. | 2020-07-14T08:27:13 | 2020-07-14T08:27:13 | {} | MEMBER |
ClickHouse | ClickHouse | 657,287,867 | 8,941 | alexey-milovidov | > 2. toString(toDateTime('2020-01-01 12:00:00', 'Unspecified')) displays 2020-01-01 12:00:00 no matter what local timezone is.
It already works this way regardless to the timezone. | 2020-07-12T22:58:52 | 2020-07-12T22:58:52 | {} | MEMBER |
ClickHouse | ClickHouse | 657,665,639 | 12,081 | alexey-milovidov | 20.6 has become prestable. If no more issues will be found, it will be released to stable this week. | 2020-07-13T16:38:52 | 2020-07-13T16:38:52 | {} | MEMBER |
ClickHouse | ClickHouse | 657,682,451 | 10,572 | alexey-milovidov | @zhang2014
Yes, they are constant expressions. | 2020-07-13T17:12:06 | 2020-07-13T17:12:06 | {} | MEMBER |
ClickHouse | ClickHouse | 657,956,567 | 12,454 | MrSsunlight | > duplicate #10994
>
> It's broken in 20.4 and 20.5 you can use 20.3.
>
> it has been fixed in testing 20.6.1.4050+
thanks | 2020-07-14T04:17:28 | 2020-07-14T04:17:28 | {} | NONE |
ClickHouse | ClickHouse | 657,997,095 | 12,443 | alexey-milovidov | @akuzm Performance test showed "2 faster, 20 slower" in queries without related code paths. | 2020-07-14T06:32:57 | 2020-07-14T06:32:57 | {} | MEMBER |
ClickHouse | ClickHouse | 658,000,235 | 12,452 | alexey-milovidov | @nikitamikhaylov I did not notice that you already have another pull request. | 2020-07-14T06:41:32 | 2020-07-14T06:41:32 | {} | MEMBER |
ClickHouse | ClickHouse | 657,281,235 | 12,424 | azat | >which just sends the query and this is not enough for this setting.
The problem is in `ParserQueryWithOutput`, when it is used (i.e. when `FORMAT` is passed) then `SETTINGS` **after `FORMAT`** are not applied:
```
$ curl http://localhost:8123/ --data-binary "SELECT DISTINCT blockSize() FROM numbers(1000000) SET... | 2020-07-12T21:56:14 | 2020-07-12T21:58:49 | {} | MEMBER |
ClickHouse | ClickHouse | 657,287,976 | 8,941 | alexey-milovidov | > 3. toTimeZone(toDateTime('2020-01-01 12:00:00'), 'Unspecified'), toTimeZone(toDateTime('2020-01-01 12:00:00', 'UTC'), 'Unspecified'), toTimeZone(toDateTime('2020-01-01 12:00:00', 'Europe/Moscow'), 'Unspecified') should give the same DateTime('Unspecified') value. From the user's point of view, it simply discards the ... | 2020-07-12T23:00:10 | 2020-07-12T23:00:10 | {} | MEMBER |
ClickHouse | ClickHouse | 657,291,392 | 12,434 | excitoon | Experienced 3x performance boost on Q1 at ontime dataset with compact storage. | 2020-07-12T23:27:33 | 2020-07-12T23:27:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,317,429 | 12,435 | l1t1 | it seems the docker/centos:8 causes the problem, I use an older version that works fine on centos:7.6, same error
```
[root@VM_0_13_centos ch206]# docker run -ti -v /tmp/ch206:/tmp centos:8 /bin/bash
[root@56d06924f336 /]# cd /tmp
[root@56d06924f336 tmp]# rpm -ivh clickhouse-common-static-20.6.1.4050-2.x86_64.rpm
... | 2020-07-13T01:51:19 | 2020-07-13T01:51:19 | {} | NONE |
ClickHouse | ClickHouse | 657,326,116 | 12,441 | alexey-milovidov | Ok, tests have passed. But they don't validate this change. | 2020-07-13T02:30:57 | 2020-07-13T02:30:57 | {} | MEMBER |
ClickHouse | ClickHouse | 657,530,856 | 12,454 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/10994
It's broken in 20.4 and 20.5 you can use 20.3.
it has been fixed in testing 20.6.1.4050+
| 2020-07-13T12:26:41 | 2020-07-13T12:26:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,606,389 | 12,468 | 4ertus2 | `set multiple_joins_rewriter_version = 2` as workaround | 2020-07-13T14:50:43 | 2020-07-13T14:50:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,043,738 | 10,561 | alesapin | Memory stress:
```
2020.07.14 08:33:10.377725 [ 51650 ] {} <Fatal> BaseDaemon: ########################################
2020.07.14 08:33:10.377864 [ 51650 ] {} <Fatal> BaseDaemon: (version 20.4.7.67 (official build)) (from thread 4789) (query_id: 16f55d4c-af63-4a84-8be1-6179741904b5) Received signal Aborted (6).
20... | 2020-07-14T08:22:57 | 2020-07-14T08:22:57 | {} | MEMBER |
ClickHouse | ClickHouse | 658,045,774 | 10,561 | alesapin | Stress tests with address sanitizer:
```
==273==ERROR: AddressSanitizer: allocator is out of memory trying to allocate 0x20000 bytes
==273==ERROR: AddressSanitizer failed to allocate 0x2000 (8192) bytes of InternalMmapVector (error code: 12)
```
It's ok, we have fixed this error in master, but I don't remember whe... | 2020-07-14T08:27:00 | 2020-07-14T08:27:00 | {} | MEMBER |
ClickHouse | ClickHouse | 658,051,159 | 12,342 | alexey-milovidov | It was broken here: https://github.com/ClickHouse/ClickHouse/pull/10758
Serialization format was changed for `avgWeighted` but in fact is was also changed for `avg` (I have missed that).
Do not upgrade before we will come up with the solution. | 2020-07-14T08:36:54 | 2020-07-14T08:47:31 | {} | MEMBER |
ClickHouse | ClickHouse | 657,351,019 | 6,005 | alexey-milovidov | @richardartoul
> My understanding (please correct me if I'm wrong) is that the storage engines in clickhouse can basically do whatever they want but when a query is executed they're responsible for generating "native format" blocks for the query execution engine to operate on.
True.
> My understanding of the ... | 2020-07-13T04:25:30 | 2020-07-13T04:26:29 | {} | MEMBER |
ClickHouse | ClickHouse | 657,587,998 | 12,386 | fastio | @alexey-milovidov Thank you for review.
> We can lower the code size further by simply registering StorageS3 and TableFunctionS3 under two different names (without implementing a separate classes).
I have removed the classes you mentioned.
| 2020-07-13T14:17:29 | 2020-07-13T14:17:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,995,203 | 12,456 | alexey-milovidov | Sporadic build failure:
> Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart. | 2020-07-14T06:27:27 | 2020-07-14T06:27:27 | {} | MEMBER |
ClickHouse | ClickHouse | 657,531,392 | 10,994 | den-crane | @MrSsunlight fixed in 20.6.1.4050+ | 2020-07-13T12:27:58 | 2020-07-13T12:27:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,269,615 | 9,115 | 4ertus2 | You've made something very strange.
1. ALTER UPDATE is special operation. Do not use it in general ETL logic.
2. Engine JOIN is a kind of optimisation of JOINs with joinGet extension. Do not use it for JOINs if general JOIN doesn't work. Do not use it in dictionary-like scenario if dictGet doesn't work.
3. It's no... | 2020-02-14T12:33:46 | 2020-07-13T13:03:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,719,206 | 12,448 | azat | >The most convenient it to create a new branch, then pull all commits from this PR, then add more commits.
@alexey-milovidov #12477 | 2020-07-13T18:27:47 | 2020-07-13T18:27:47 | {} | MEMBER |
ClickHouse | ClickHouse | 658,000,663 | 12,438 | alexey-milovidov | Results of performance test has been changed exactly as in https://github.com/ClickHouse/ClickHouse/pull/12443 | 2020-07-14T06:42:46 | 2020-07-14T06:42:46 | {} | MEMBER |
ClickHouse | ClickHouse | 657,311,665 | 12,435 | l1t1 | I download the newer version and the same error occurs
```
[root@efe6e218fb64 /]# cd /tmp
[root@efe6e218fb64 tmp]# rpm -ivh clickhouse-server-20.6.1.4062-2.noarch.rpm
warning: clickhouse-server-20.6.1.4062-2.noarch.rpm: Header V4 RSA/SHA1 Signature, key ID e0c56bd4: NOKEY
error: Failed dependencies:
clic... | 2020-07-13T01:24:30 | 2020-07-13T01:24:30 | {} | NONE |
ClickHouse | ClickHouse | 657,652,128 | 9,887 | vkingnew | this functions are very import to analyse complex query,desire it!
>CTE syntax is different
CTE returning a table is not supported
Window functions are not supported
RANK OVER is not supported
>MySQL 8.0 has implements windows functions:
https://dev.mysql.com/doc/refman/8.0/en/window-function-descriptions.html | 2020-07-13T16:12:37 | 2020-07-13T16:14:30 | {} | NONE |
ClickHouse | ClickHouse | 657,997,570 | 12,443 | alexey-milovidov | > Yandex synchronization check (only for Yandex employees) Pending — error, check will be restarted
False claim, the check was not restarted, CC @exprmntr. | 2020-07-14T06:34:17 | 2020-07-14T06:34:17 | {} | MEMBER |
ClickHouse | ClickHouse | 657,278,030 | 10,957 | vitlibar | Currently (after PR https://github.com/ClickHouse/ClickHouse/pull/11670) `ON CLUSTER` can be written no matter where: after policy's name or after `ON db.table`, it works. In the older versions you had to write `ON CLUSTER` after `ON db.table`, as you said. | 2020-07-12T21:29:26 | 2020-07-12T21:29:58 | {} | MEMBER |
ClickHouse | ClickHouse | 657,286,484 | 12,405 | alexey-milovidov | CC @Enmk | 2020-07-12T22:45:59 | 2020-07-12T22:45:59 | {} | MEMBER |
ClickHouse | ClickHouse | 657,294,860 | 12,428 | alexey-milovidov | Most likely will be closed. | 2020-07-12T23:54:30 | 2020-07-12T23:54:30 | {} | MEMBER |
ClickHouse | ClickHouse | 657,313,003 | 12,438 | alexey-milovidov | @akuzm The result of performance test is not published. | 2020-07-13T01:30:18 | 2020-07-13T01:30:18 | {} | MEMBER |
ClickHouse | ClickHouse | 657,326,333 | 12,424 | alexey-milovidov | > Performance — 15 faster, 215 unstable
Wow, it has passed. | 2020-07-13T02:32:01 | 2020-07-13T02:32:01 | {} | MEMBER |
ClickHouse | ClickHouse | 657,358,387 | 12,081 | alexey-milovidov | About this task - it is a checklist for significant issues that must be fixed before the next release, not for features.
New features come to whatever release that is going to be next after the feature is merged to master. | 2020-07-13T04:58:10 | 2020-07-13T04:58:10 | {} | MEMBER |
ClickHouse | ClickHouse | 657,358,796 | 12,081 | alexey-milovidov | We consolidate issues according to the following criteria:
- regressions (something works worse or with unintended broken compatibility with previous versions);
- major issues (e.g. reproducible crashes). | 2020-07-13T04:59:36 | 2020-07-13T04:59:36 | {} | MEMBER |
ClickHouse | ClickHouse | 657,427,114 | 12,455 | amosbird | @zlobober btw, your NULL_LAST logic has gone (I don't know why). I've added it back https://github.com/ClickHouse/ClickHouse/pull/12455/files#diff-043ede2be3d33df96a1f9b9b355678ebL89 | 2020-07-13T09:06:37 | 2020-07-13T09:06:37 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 657,535,845 | 11,642 | abyss7 | @bookeezhou Do you build on a native ARM platform? What is your distribution? What is your cmake configuration command-line? | 2020-07-13T12:38:06 | 2020-07-13T12:38:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,649,044 | 4,190 | akuzm | > @akuzm, could you help to draw up the detailed requirements for the necessary contribution?
> I'll try to find someone and pay for his work ([ru link](https://freelance.habr.com/tasks/304258))
Just implement what you need and send a PR, we don't have any particular requirements other than working code + docs + te... | 2020-07-13T16:06:49 | 2020-07-13T16:06:49 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,667,097 | 8,062 | alexey-milovidov | > миллион раз сделать лукап (nested-loop) по индексу
ClickHouse не делает lookups в цикле для работы индекса. | 2020-07-13T16:41:44 | 2020-07-13T16:41:44 | {} | MEMBER |
ClickHouse | ClickHouse | 657,679,608 | 12,278 | alexey-milovidov | @akuzm This PR does not affect `size`, only `capacity`.
I can only suspect cache alignment or differences in false sharing but there are no differences in perf events metrics. | 2020-07-13T17:06:19 | 2020-07-13T17:06:19 | {} | MEMBER |
ClickHouse | ClickHouse | 657,946,598 | 12,478 | RajatGoyal | @zhang2014 we could have a syntax like mysql table function which would accept an engine specific query, where you can't use CH functions.
```
For example:
select b.name, a. value from ch.table a
inner join on_engine.mysql_database ( external database compatible query ) b
```
Will check out MaterializeMySQL... | 2020-07-14T03:34:24 | 2020-07-14T04:53:36 | {} | NONE |
ClickHouse | ClickHouse | 658,046,457 | 10,561 | alesapin | ClickHouse Unit Tests Msan -- infinite recursion in libunwind. | 2020-07-14T08:28:16 | 2020-07-14T08:28:16 | {} | MEMBER |
ClickHouse | ClickHouse | 657,288,610 | 8,942 | alexey-milovidov | It can be also named `Calendar` and represent calendar date with offset from UTC as in ISO-8601:
`2020-07-13T02:04:03.123+0300`. | 2020-07-12T23:05:00 | 2020-07-12T23:05:00 | {} | MEMBER |
ClickHouse | ClickHouse | 657,416,748 | 12,448 | azat | >CC @azat Look at the warnings from Arrow.
Ok | 2020-07-13T08:48:32 | 2020-07-13T08:48:32 | {} | MEMBER |
ClickHouse | ClickHouse | 657,617,363 | 10,572 | zhang2014 | Why ```select touint32 (25703952), todatetime ('2014-03-19 23:59:58 ')``` return const column, this is expected behavior?
cc : @alexey-milovidov | 2020-07-13T15:09:48 | 2020-07-13T15:10:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 657,821,532 | 12,424 | azat | > > AFAIU this is the intention, but looks tricky, maybe worth adding them to the ASTSelectQuery too after parsing from ParserQueryWithOutput?
>
> I don't remember any intentions about that.
@alexey-milovidov so here is an RFC - #12480 (at least it will run tests and perf-tests with SETTINGS after FORMAT interpret... | 2020-07-13T22:18:51 | 2020-07-13T22:18:51 | {} | MEMBER |
ClickHouse | ClickHouse | 657,933,468 | 12,478 | zhang2014 | > The request is for a feature to force a query to run completely on MySQL as is and return results back to CH, without CH trying to parse the query and try to get all data from every table separately in the system.
This is difficult, for example how we execute functions that don't exists in MySQL is a problem(curre... | 2020-07-14T02:44:45 | 2020-07-14T02:44:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,004,436 | 12,424 | alexey-milovidov | BTW, Cumulative graph is great. | 2020-07-14T06:52:51 | 2020-07-14T06:52:51 | {} | MEMBER |
ClickHouse | ClickHouse | 658,045,175 | 10,561 | alesapin | Functional tests with memory sanitizer -- `01081_demangle`. Strange error, but it works in the release build, so skipping it. | 2020-07-14T08:25:50 | 2020-07-14T08:26:01 | {} | MEMBER |
ClickHouse | ClickHouse | 657,303,691 | 12,433 | amosbird | Irrelevant failure due to test is flaky. | 2020-07-13T00:47:26 | 2020-07-13T00:47:26 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 657,440,126 | 7,435 | CurtizJ | The data is deleted, but only empty parts are remaining. Duplicate of #5491. | 2020-07-13T09:27:32 | 2020-07-13T09:27:32 | {} | MEMBER |
ClickHouse | ClickHouse | 657,553,637 | 10,561 | alesapin | All important tests should be fixed, but flaky failures still may happen. | 2020-07-13T13:15:06 | 2020-07-13T13:15:06 | {} | MEMBER |
ClickHouse | ClickHouse | 657,303,768 | 12,418 | amosbird | Irrelevant failure due to test is flaky. | 2020-07-13T00:47:47 | 2020-07-13T00:47:47 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 657,313,853 | 12,448 | alexey-milovidov | Sandbox is totally broken:
```
2020-07-13 04:19:44
Unhandled exception: Traceback (most recent call last):
File "sandbox/bin/executor.py", line 1043, in _execute
self.task.on_execute()
File "sandbox/projects/clickhouse/BaseOnCommitTask/base.py", line 295, in on_execute
state, description, url = self.... | 2020-07-13T01:34:01 | 2020-07-13T01:34:01 | {} | MEMBER |
ClickHouse | ClickHouse | 657,330,536 | 10,572 | alexey-milovidov | It's related to `bloom_filter` index, CC @zhang2014. | 2020-07-13T02:50:34 | 2020-07-13T02:50:34 | {} | MEMBER |
ClickHouse | ClickHouse | 657,462,007 | 12,450 | alesapin | Can you provide full logs around this exception? | 2020-07-13T10:03:02 | 2020-07-13T10:03:12 | {} | MEMBER |
ClickHouse | ClickHouse | 657,693,508 | 12,448 | azat | >Need to fix the release build.
These are -Wdeprecated from arrow, to fix them I need some bits from this PR (`s/PUBLIC/PRIVATE/` + other rabbit mq bits), so I will pick then and fix deprecated API usage for arrow on top of it, and submit it as a separate PR
Or I can just send a patch on top of this PR here, @ale... | 2020-07-13T17:34:39 | 2020-07-13T17:34:39 | {} | MEMBER |
ClickHouse | ClickHouse | 658,003,144 | 12,424 | alexey-milovidov | Example: https://github.com/ClickHouse/ClickHouse/pull/8311 | 2020-07-14T06:49:29 | 2020-07-14T06:49:29 | {} | MEMBER |
ClickHouse | ClickHouse | 658,339,506 | 9,882 | alexey-milovidov | Fixed in #9890 | 2020-07-14T18:26:45 | 2020-07-14T18:26:45 | {} | MEMBER |
ClickHouse | ClickHouse | 658,491,741 | 10,429 | alexey-milovidov | @l1t1 BTW, there is related task https://github.com/ClickHouse/ClickHouse/issues/12414
But I'm not sure about use cases for it.
What is your use case? Maybe you are using some system with limited terminal capabilities? | 2020-07-15T01:23:14 | 2020-07-15T01:23:14 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.